High efficiency saturated amplifiers and switching power amplifiers are known in the art. One drawback with the use of such amplifiers is the high peak voltages relative to the dc supply that the active devices must withstand in these modes of operation. In order to improve the gain, switching speed, and on-resistance of transistors, the breakdown voltage of the device is usually reduced. This tradeoff is exhibited by all modern semiconductor device technologies, including but not limited to field-effect transistors (FET), bipolar junction transistors (BJT), heterojunction bipolar transistors (HBT), high electron mobility transistors (HEMT), metal-semiconductor field-effect transistors (MESFET), metal-oxide semiconductor field effect transistors (MOSFET), and junction field-effect transistors (JFET). The effect is also independent of the semiconductor material system from which the devices are constructed, including but not limited to gallium arsenide (GaAs), indium phosphide (InP), silicon-germanium (SiGe), and silicon (Si) processes such as silicon bipolar (Si BJT), complementary metal oxide field effect transistor (CMOS) processes, and silicon-on-insulator (SOI) technologies.
In high efficiency switching amplifiers, such a reduction in breakdown voltage can be problematic. Unlike many applications in which the maximum voltage seen by any device is typically limited to the dc voltage of the power source, high efficiency switching amplifiers such as class E, class F, class inverse-F, current-mode class D and class E/F can require that the peak voltage seen by the devices be several times the dc supply. Class F, for instance, can require a peak voltage at least twice the supply voltage, whereas class E can require the device to withstand over 3.5 times the supply voltage without breaking down.
This high peak voltage relative to the dc power supply voltage applied results from the use of an inductor to connect the active device to the dc supply voltage. FIG. 1 is a diagram of a generalized circuit topology typically used in saturated and switching amplifiers such as class E, class F, and class E/F. The active device is connected to the dc supply through the inductor. Since the dc (or average) voltage drop across any inductor at steady state can be zero, the voltage waveform can have an average voltage equal to the supply voltage. This corresponds to a limitation on the waveform that the average area above the supply voltage and the area below it must be the same. This can be seen in FIG. 2, depicting typical waveforms for a representative switching amplifier, with equal areas above and below the supply voltage shaded.
As can be seen in FIG. 2, the active device spends a significant portion of its time in a low voltage state. This is so that the active device can conduct the bulk of its current during this time, thereby reducing the power dissipation in the device, resulting in high efficiency. Unfortunately, this results in a very large area below the supply voltage, necessitating an equally large area above it. Thus the voltage during the times when the switch is not low can be significantly greater than the supply voltage, usually by a factor of two to four.
In a typical CMOS process, for instance, the device breakdown can be less than 6 V whereas the supply voltage is in many cases 3.3 V or higher. With a 3.3 V supply, the class E amplifier can produce waveforms with peak voltage greater than 11V, almost twice that which a CMOS device with 6 V breakdown can tolerate. Thus in this application, the supply voltage can be changed, a more expensive high-voltage process can be used, or a less efficient type of power amplifier with a lower peak voltage can be employed. If the supply voltage cannot be changed, such as if it is coming from a battery or if other circuits on the same supply cannot change their supply voltage, the high peak to supply ratio of the traditional switching amplifiers thus forces a sacrifice in either cost or performance.